Melos Maestro modification log (updated Oct 05, 2008) - LARGE PHOTOS!

[nikongod]

Quote:

Originally Posted by fzman /img/forum/go_quote.gif paralleling two identical caps in the signal path may not work out quite how you want it to -- it may cause a subtle blur that using dissimilar values in parallel may not create -- if you have the resources, try it both ways before you spring for the $$$ caps. just my 2 cents worth....

From my experience that is backwards.

Paralleling caps of dis-similar values causes the blur.

 

[dsavitsk]

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif My top choice for bang/buck is the Sonicap Gen 1 1.0uF/200V

This is what I used, too.

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif I'm also looking at changing the pre-amp output caps and really don't see why they would be rated at 250V. Looking at the schematic there is a 10uF cap bypassed by 0.1uF. I'm thinking Sonicap for both - Gen1 10.0uF/200V bypassed by Gen2 0.1uF/200V - as long as 200V is enough.

I think you ar right that there is no voltage on them as the pre outs (at least on my SHA-1) are right after the phone outs. The reason there are caps is that the amp can have a small amount of offset. This is not problem for headphones, but since the next stage could amplify it, it is standard to put coupling caps to prevent that possibility. They probably are rated high voltage because nobody makes low voltage film caps.

The real question is why you would want 10uF, or a bypass. If you won't use this as a preamp, skip these. If you will, figure out what the next stage Zin is, and replace them appropriatly.

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif The cap which goes from the grid of the tube to ground (the other one should connect to signal input) is of questionable importance (quality wise). It is necessary as part of the phase-splitter, but is not DIRECTLY in the signal path.

To the contrary, they are directly in the signal path and matter as much as any others.


Quote:

Paralleling caps of dis-similar values causes the blur.

I agree with this. Indeed, paralleling same values sometimes can improve things and is why Multicaps work. I have tried paralleling Vitamin Q's, for instance, with very good luck.

 

[nikongod]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif I think you ar right that there is no voltage on them as the pre outs (at least on my SHA-1) are right after the phone outs. The reason there are caps is that the amp can have a small amount of offset. This is not problem for headphones, but since the next stage could amplify it, it is standard to put coupling caps to prevent that possibility. They probably are rated high voltage because nobody makes low voltage film caps.

The preamp outs on the SHA-gold are set up differently than on the SHA-1.

The gold runs the mosfets for the preamp from the high voltage supply, they are totally separate from the headphone amp. The caps should be blocking some wicked DC offset here.
The SHA-1 uses the headphone driver for its preamp out. There should be virtually no offset.
Quote:

To the contrary, they are directly in the signal path and matter as much as any others.

They are just there to hook the non-input side of the long-tail-pair to actual signal ground. Does the signal "reflect" through the grid of the "non-driven' side?

 

[dsavitsk]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif The preamp outs on the SHA-gold are set up differently than on the SHA-1.

Interesting

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif They are just there to hook the non-input side of the long-tail-pair to actual signal ground. Does the signal "reflect" through the grid of the "non-driven' side?

Signal travels in a loop, not a line, so you need to account for the return signal. A cap connection allows AC and blocks DC, or indeed any DC reference. The only reason to cap couple something is because you want signal to go through it. Try disconnecting it and see what happens. Or, try connecting it with a 10pF cap and see where the frequency response goes.

 

[FallenAngel]

Great discussion here guys - to be safe, I'll order another 1uF/200V Sonicap Gen1 (I really love the price for this size, $5.80 is a steal!) for that position, getting very excited now.

 

[nikongod]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif Signal travels in a loop, not a line, so you need to account for the return signal. A cap connection allows AC and blocks DC, or indeed any DC reference. The only reason to cap couple something is because you want signal to go through it. Try disconnecting it and see what happens. Or, try connecting it with a 10pF cap and see where the frequency response goes.

So I tried this, and at the very least agree that the caps cant be removed from the circuit (even without jumpering, just leaving them out).

I disconnected one of the "non-driven side" caps from ground, and let her rip. I was expecting at worst a few db drop on that channel, but got no music, and a decent bit of static which did not change with volume knob position. I did not try to replace the cap with a small ceramic.

It is clearly the case that the caps are necessary even in this position, and probably passing the signal, but I dont understand why. The caps are obviously blocking DC between the grid and ground, but the grids have leak resistors. Why dont the leak resistors provide enough reference for the tube to run properly?

I suppose I will cede defeat in this debate (although I have gained knowledge to improve my gear: I win!) but would GREATLY appreciate any information on why this works the way it does.

on the note of the preamp caps, and stuff:
I measured my preamp outs, the caps are holding back about 80VDC.
I would say that the 10uF caps are necessary as a GENERAL design item: a single ended amp with a 10k-ohm input impedance needs them to keep all effects of the low-frequency roll off out of the audible range. They are overkill for the balanced outputs, but since they should both be the same... you can use less if you dont want to drive an amp with a 10k-ohm input impedance.

 

[FallenAngel]

Very cool.

6x Sonicap Gen1 on the way for coupling and the Grid-Ground position. Too bad I can't fit the Vitamin Q in there.

The other 1uF caps look like they're for the PSU - any suggestions for those? I'm looking at some normal quality BC MKT but is it even worth it?

 

[dsavitsk]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif but would GREATLY appreciate any information on why this works the way it does.

The short answer is that the second grid is AC coupled to signal ground. Otherwise, what would that side of the tube be doing? The only real difference between balanced and single ended signals is that in single ended one side of each is tied to ground, that is, they are referenced to the same thing. In balanced they are separate. So, since this is splitting the signal, you need the second side of the signal to come from somewhere. Look at the Schmidt splitter here.

A better link: http://www.freewebs.com/valvewizard/acltp.html

 

[FallenAngel]

Well, I've been BUSY today! I finished a Dynalo and a custom PSU for it, just needs casing, then finished building a MiniMAX prototype which is also awaiting casing, then I destroyed the unbalanced outputs of my ESI Juli@ with a bad opamp replacement, hopefully the balanced still work (opamps also replaced with AD8022 in MSOP packages) and then I started on the Melos. Whew.

So the replacement caps came in today. The values are from the Melos SHA-Gold schematic.

Low-Voltage caps are Nichicon UHE and one Panasonic FM
High-Voltage caps are Panasonic TS-HA except I took nikongod's recommendation about the 100uF/160V cap and replaced that with a 100uF/250V AXON (Solen) cap. I don't know about it "sounding like sex", but the amp does sound pretty damn good. I'll wait for everything to burn in for a while before passing judgment but so far it's very nice and dead quiet in terms of background noise.

One weird thing I noticed is that the old caps I took out are so much lighter than the new ones I'm putting in. They feel like paper compared to the new ones.

Finally replaced high-voltage diode bridge to HEXFREDs.

I also took the time to reroute the headphone jack wiring more "directly" instead of around the entire case.

Ok, now it's almost 4am and I have to go to work tomorrow.

I'm going to sleep.

 

[nikongod]

very nice indeed

A capacitors capacitance is totally over-rated when dealing with tube power supplies. Quality>>quantity.

Although its more work (hehe there is always more work) I would move the film cap.

The cap near the "M" in melos is should probably be the film. The one where it is now is before the voltage regulator. The one near the M feeds the tubes. Swapping them should do amazing things for the sound.

When I moved parts from my SHA-1 back to my Gold, I didnt put the film cap back. I miss it.

 

[FallenAngel]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif very nice indeed A capacitors capacitance is totally over-rated when dealing with tube power supplies. Quality>>quantity. Although its more work (hehe there is always more work) I would move the film cap. The cap near the "M" in melos is should probably be the film. The one where it is now is before the voltage regulator. The one near the M feeds the tubes. Swapping them should do amazing things for the sound. When I moved parts from my SHA-1 back to my Gold, I didnt put the film cap back. I miss it.

Are you sure that 100uF is enough? Schematic shows 1000uF for that position and if I just swap them,Ii'll be swapping 1000uF and 100uF?! Sounds a little dangerous.

 

[FallenAngel]

With the completion of a couple of cable projects and a Dynalo, I finally got some time to work on the Melos.

Taking nikongod's advice about moving the cap huge film cap, I have rearranged the caps a bit.

1000uF TSHA is now 100uF Solen
470uF TSHA is now 1000uF TSHA
100uF Solen is now 470uf TSHA

I figured that would increase capacitance by quite a bit, but no 10x in one place.

Since I got the Sonicap 1uF, they were also installed. 4x for inter-stage coupling and 2x for Grid to Ground (opposite side of signal).

Well, I if I might say so myself, it sounds absolutely incredible! The bass is through the roof with the RS-1 and the detail is incredible. By far the best I've ever heard the RS-1 sound. I prefer this to the Beta22, screw neutrality, RS-1 is mean to be FUN!

Next up, I don't even know yet.

Probably will swap out some RCA to nicer ones and perhaps rewire the amp. I like this amp so much as a headphone amp, I'm seriously considering removing the source selector and rewiring it as single-input, skip the source-selector and optimize the wiring. That's for another day though, this amp is getting closed for another day, it's been through enough that I will just enjoy it.

 

[mwofsi]

Nice work FallenAngel, glad it's turned out so well. That Sonicap arrangement between the valves is great. Any chance of a pic or two of the whole thing?

 

[FallenAngel]

Thanks. I'll take a full pic later on, it's finally closed for the first time in over a month and I like it that way. To get a good impression of it though, just take a look at the last two photos and combine them .

I also realized I might not last too long before going forward with another upgrade I've been thinking about, the PSU film caps. What would be a good option for these? I'm thinking the Wima MKP4 would be a good choice but I don't have any on hand (although that can change very quickly); I do have some BC components MKT and MKP caps that I can use. Any opinions?

Thanks in advance.

 

[FallenAngel]

PSU film bypass recommendations anybody?

I'm going to place an order with Mouser very soon, will likely go with MKP10, same as the Millet Hybrid MAX design. Just have to make sure to measure size.